Topic Monitoring

View live, aggregated topic data about publications, subscribers, messages, and publishers.

Topics are not monitored automatically because maintaining statistics for many topic monitors might affect the performance of IBM® IoT MessageSight. You must define specific topic monitors in order to get information about those monitored subtrees. Predefined queries can be run to view the following metrics:

• Topics with Most Messages Published

  This query shows the monitored subtrees that are published to the most. Use this statistic to understand how much traffic a topic is producing. A higher than expected figure might indicate that there is an issue that requires further investigation.

• Topics with Most Subscribers

  This query shows the monitored subtrees that are subscribed to the most. Use this information to understand how many unique subscribers there are for a particular topic. If you have many subscribers to a particular topic, you might want to consider subdividing the topic string to make it more precise.

• Topics with Most Messages Rejected

  This query shows the monitored subtrees with the highest occurrence of a publication which could not be buffered for 1 or more subscribers. The publish completed because the quality of service allowed it to complete. This statistic might indicate that the queue depth is at the maximum limit, and as a result no more messages can be published. Investigation into why the messages are not being removed from the queue might be required. For example, you might consider changing a messaging policy to increase the maximum messages of subscribers in a subtree, or changing subscribing applications so that they consume messages at a faster rate.

• Topics with Most Failed Publishes

  This query shows the monitored subtrees that have the highest occurrence of publications which failed. This statistic might indicate that publishers are publishing to an incorrect topic string, or publishing to topics for which they are not authorized. Alternatively, the failure might be because a subscriber did not buffer a message with a high quality of service. You might want to consider changing a messaging policy to increase the maximum number of messages that subscribers can buffer.

• Topics with Least Messages Published

  This query shows how much traffic a topic is producing. A lower than expected figure might indicate that there is an error. If a topic has zero publications, then you might want to investigate why no traffic is generated on that topic.

• Topics with Least Subscribers

  This query shows the monitored subtrees that are subscribed to the least. Use this information to understand how many subscribers there are for a particular topic. If a topic has no subscribers, then you might want to stop publishing to that topic in future. Alternatively, it might be that the topic is publishing an incorrect string, or that subscribers are subscribing to an incorrect string.

• Topics with Least Messages Rejected

  This query shows the monitored subtrees that have the lowest occurrence of messages which failed. This statistic might indicate that the maximum queue depth can be reduced, or provide an indication that the topic is in a healthy state.

• Topics with Least Failed Publishes

  This query shows the monitored subtrees that have the lowest occurrence of publications which failed. This statistic might indicate that publishers have the appropriate authorization, or that there are little or no problems with incorrect strings. Any specific publishes which are rejected, can be investigated.